From 5e3c68b62dd72255db95c5822835a3931770f285 Mon Sep 17 00:00:00 2001
From: Chris Dyer <cdyer@cs.cmu.edu>
Date: Tue, 12 Jul 2011 22:34:34 -0400
Subject: debugged pro trainer

---
 utils/filelib.h | 1 +
 1 file changed, 1 insertion(+)

(limited to 'utils/filelib.h')

diff --git a/utils/filelib.h b/utils/filelib.h
index dda98671..a8622246 100644
--- a/utils/filelib.h
+++ b/utils/filelib.h
@@ -12,6 +12,7 @@
 
 bool FileExists(const std::string& file_name);
 bool DirectoryExists(const std::string& dir_name);
+void MkDirP(const std::string& dir_name);
 
 // reads from standard in if filename is -
 // uncompresses if file ends with .gz
-- 
cgit v1.2.3


From ed2b1ce7181fb21d2363b0d5ce04d7fa52aef0fa Mon Sep 17 00:00:00 2001
From: Chris Dyer <cdyer@cs.cmu.edu>
Date: Sat, 17 Sep 2011 01:08:45 +0100
Subject: enable ramdisk scratch for per-sentence-grammars

---
 training/mpi_batch_optimize.cc | 35 +++++++++++++++++++++++++++++++++++
 utils/filelib.cc               | 19 +++++++++++++++++++
 utils/filelib.h                |  5 +----
 3 files changed, 55 insertions(+), 4 deletions(-)

(limited to 'utils/filelib.h')

diff --git a/training/mpi_batch_optimize.cc b/training/mpi_batch_optimize.cc
index cc5953f6..0ba8c530 100644
--- a/training/mpi_batch_optimize.cc
+++ b/training/mpi_batch_optimize.cc
@@ -22,6 +22,7 @@ namespace mpi = boost::mpi;
 #include "ff_register.h"
 #include "decoder.h"
 #include "filelib.h"
+#include "stringlib.h"
 #include "optimize.h"
 #include "fdict.h"
 #include "weights.h"
@@ -42,6 +43,7 @@ bool InitCommandLine(int argc, char** argv, po::variables_map* conf) {
 	("correction_buffers,M", po::value<int>()->default_value(10), "Number of gradients for LBFGS to maintain in memory")
         ("gaussian_prior,p","Use a Gaussian prior on the weights")
         ("means,u", po::value<string>(), "File containing the means for Gaussian prior")
+        ("per_sentence_grammar_scratch,P", po::value<string>(), "(Optional) location of scratch space to copy per-sentence grammars for fast access, useful if a RAM disk is available")
         ("sigma_squared", po::value<double>()->default_value(1.0), "Sigma squared term for spherical Gaussian prior");
   po::options_description clo("Command line options");
   clo.add_options()
@@ -186,6 +188,36 @@ struct VectorPlus : public binary_function<vector<T>, vector<T>, vector<T> >  {
   } 
 }; 
 
+void MovePerSentenceGrammars(const string& root, int size, int rank, vector<string>* c) {
+  if (!DirectoryExists(root)) {
+    cerr << "Can't find scratch space at " << root << endl;
+    abort();
+  }
+  ostringstream os;
+  os << root << "/psg." << size << "_of_" << rank;
+  const string path = os.str();
+  MkDirP(path);
+  string sent;
+  map<string, string> attr;
+  for (unsigned i = 0; i < c->size(); ++i) {
+    sent = (*c)[i];
+    attr.clear();
+    ProcessAndStripSGML(&sent, &attr);
+    map<string, string>::iterator it = attr.find("grammar");
+    if (it != attr.end()) {
+      string src_file = it->second;
+      bool is_gzipped = (src_file.size() > 3) && (src_file.rfind(".gz") == (src_file.size() - 3));
+      string new_name = path + "/" + md5(sent);
+      if (is_gzipped) new_name += ".gz";
+      CopyFile(src_file, new_name);
+      it->second = new_name;
+    }
+    ostringstream ns;
+    ns << SGMLOpenSegTag(attr) << ' ' << sent << " </seg>";
+    (*c)[i] = ns.str();
+  }
+}
+
 int main(int argc, char** argv) {
 #ifdef HAVE_MPI
   mpi::environment env(argc, argv);
@@ -257,6 +289,9 @@ int main(int argc, char** argv) {
   ReadTrainingCorpus(conf["training_data"].as<string>(), rank, size, &corpus);
   assert(corpus.size() > 0);
 
+  if (conf.count("per_sentence_grammar_scratch"))
+    MovePerSentenceGrammars(conf["per_sentence_grammar_scratch"].as<string>(), rank, size, &corpus);
+
   TrainingObserver observer;
   while (!converged) {
     observer.Reset();
diff --git a/utils/filelib.cc b/utils/filelib.cc
index a0969b1a..d206fc19 100644
--- a/utils/filelib.cc
+++ b/utils/filelib.cc
@@ -2,6 +2,12 @@
 
 #include <unistd.h>
 #include <sys/stat.h>
+#include <sys/types.h>
+#include <sys/socket.h>
+#include <cstdlib>
+#include <cstdio>
+#include <sys/stat.h>
+#include <sys/types.h>
 
 using namespace std;
 
@@ -32,3 +38,16 @@ void MkDirP(const string& dir) {
   }
 }
 
+#if 0
+void CopyFile(const string& inf, const string& outf) {
+  WriteFile w(outf);
+  CopyFile(inf,*w);
+}
+#else
+void CopyFile(const string& inf, const string& outf) {
+  ofstream of(outf.c_str(), fstream::trunc|fstream::binary);
+  ifstream in(inf.c_str(), fstream::binary);
+  of << in.rdbuf();
+}
+#endif
+
diff --git a/utils/filelib.h b/utils/filelib.h
index a8622246..bb6e7415 100644
--- a/utils/filelib.h
+++ b/utils/filelib.h
@@ -113,9 +113,6 @@ inline void CopyFile(std::string const& inf,std::ostream &out) {
   CopyFile(*r,out);
 }
 
-inline void CopyFile(std::string const& inf,std::string const& outf) {
-  WriteFile w(outf);
-  CopyFile(inf,*w);
-}
+void CopyFile(std::string const& inf,std::string const& outf);
 
 #endif
-- 
cgit v1.2.3